© 2012 Nature America, Inc. All rights reserved.
nature protocols | VOL.7 NO.4 | 2012 | 771
iG model for the study of food intake and GI physiology
Direct delivery of nutrients to stomach or intestine via an
implanted catheter allows the testing of diverse and unique ques-
tions. These include the roles of a segment within the normal
food intake route in neuroendocrine mechanisms of cephalic,
GI and metabolic responses to nutrient intake1–7. Mechanisms of
inhibition of gastric emptying or postgastric satiety are studied
by using iG or intraduodenal infusion of a liquid diet in mice that
have been given receptor antagonists for intestinal or pancreatic
endocrine peptides such as cholecystokinin or amylin8,9. The use
of the iG model facilitates the identification of rapid postoral
mechanisms of flavor preference conditioning by sugar and fat
as well as its potential interaction with an orosensory response3,5.
The maintenance of a long-term access (2–3 months) to the GI
tract is often challenging. However, if it is achieved, it enables the
study of the physiologic or metabolic adaptation to and patho-
physiological consequences of experimental dietary conditioning.
Such approaches are becoming crucial to studies on physiologic
and metabolic complications associated with lifestyle-related
diseases such as metabolic syndrome associated with obesity as
Modeling liver disease using the iG model
One of the early objectives for creation of the rodent long-term iG
model was the scientific desire to test whether maximal ethanol
intake results in major alcoholic liver disease (ALD) in rodents
that are naturally averse to ethanol and otherwise do not volun-
tarily consume adequate amounts of ethanol to damage the liver.
Answers derived from the model for this critical question were both
yes and no. With a maximal intake of ethanol reaching as high as
49% of caloric intake facilitated by forced iG ethanol administra-
tion, the rat develops severe hepatic steatosis and elevated alanine
aminotranferease (ALT) levels even with a diet low in fat, which
does not cause these changes in the models with ad libitum con-
sumption of ethanol-containing liquid diet10. An increase in the
dietary content of polyunsaturated fat causes further progression
to steatohepatitis11, and the addition of carbonyl iron12 or enteral
lipopolysaccharide administration13 potentiates liver fibrosis. Thus,
although maximal ethanol intake indeed causes more serious liver
injury than that observed in other ad libitum models, it alone is not
sufficient for progression of ALD. The development of the mouse
iG model has revolutionized the field because it allows genetic stud-
ies of ALD. The iG model using mice deficient in tumor necrosis
factor (TNF) receptor I, CD14, Toll-like receptor 4, LBP, p47phox,
ICAM-1 and so on revealed the importance of endotoxin-mediated
activation of hepatic macrophages, their NADPH oxidase activa-
tion and generation of reactive oxygen species, subsequent TNF-α
induction and proinflammatory response in the ALD pathogenesis
(see ref. 14 for a review).
Application to obesity-associated research
Another important application of the mouse iG model is obesity-
associated metabolic syndrome including nonalcoholic fatty
liver disease (NAFLD), which is recognized as a global epidemic
and the most common type of liver disease15. Although NAFLD
and more pathologically advanced nonalcoholic steatohepatitis
(NASH) are associated with obesity-associated insulin resistance
(IR), the precise pathogenetic relationship between them histori-
cally could not be studied because of the lack of animal models that
reproduce the natural history and course of the disease in humans.
Genetic models such as mice deficient in leptin16 and leptin recep-
tor17 are not clinically relevant and do not develop steatohepatitis.
Mice with targeted ablation of peroxisome proliferator–activated
receptor-α (Ppara)18, acyl-CoA oxidase (Acox1)19, methionine
adenosyltransferanse1A (Mat1a)20 and phosphatase and tension
homolog deleted on chromosome 10 (Pten)21 all develop deranged
liver metabolism and consequent NAFLD or/and NASH, but
these mutations are, again, rare in patients. Feeding rodents a diet
deficient in choline and methionine results in NASH22, but the
model does not result in the development of obesity or IR, and
this extreme malnutrition is not relevant to NAFLD and NASH in
human patients. More natural nutrition models include feeding
ad libitum a diet high in sucrose23,24, fructose23 or fat24. However,
these models produce mild obesity and NAFLD but not NASH.
The most common cause of obesity in humans is hypercaloric
alimentation. The iG model can be used to reproduce this condi-
tion by forced feeding. Indeed, iG feeding of male C57BL/6 mice
with high-fat diet at a level up to 85% in excess of the standard
Mouse intragastric infusion (iG) model
Akiko Ueno1,2, Raul Lazaro1,2, Ping-Yen Wang1,2, Reiichi Higashiyama1,2, Keigo Machida1,3 & Hidekazu Tsukamoto1,2,4
1Southern California Research Center for ALPD and Cirrhosis, Keck School of Medicine of the University of Southern California, Los Angeles, California, USA.
2Department of Pathology, Keck School of Medicine of the University of Southern California, Los Angeles, California, USA. 3Department of Molecular Microbiology
and Immunology, Keck School of Medicine of the University of Southern California, Los Angeles, California, USA. 4Department of Veterans Affairs Greater Los Angeles
Healthcare System, Los Angeles, California, USA. Correspondence should be addressed to H.T. (email@example.com).
Published online 29 March 2012; doi:10.1038/nprot.2012.014
Direct intragastric delivery of a diet, nutrient or test substance can be achieved in rodents (mice and rats) on a long-term
(2–3 months) basis using a chronically implanted gastrostomy catheter and a flow-through swivel system. this rodent intragastric
infusion (iG) model has broad applications in research on food intake, gastrointestinal (GI) physiology, GI neuroendocrinology,
drug metabolism and toxicity, obesity and liver disease. It achieves maximal control over the rate and pattern of delivery and it can
be combined with normal ad libitum feeding of solid diet if so desired. It may be adopted to achieve infusion at other sites of the
GI system to test the role of a bypassed GI segment in neuroendocrine physiology, and its use in genetic mouse models facilitates
the genetic analysis of a central question under investigation.
© 2012 Nature America, Inc. All rights reserved.
772 | VOL.7 NO.4 | 2012 | nature protocols
intake spontaneously produces severe obesity, abdominal adipos-
ity, hyperglycemia, hyperinsulinemia, hyperleptinemia, glucose
tolerance, IR and NAFLD/NASH25. A model such as this would
allow studies on interrelationships among these metabolic and
pathologic phenotypes in their pathogenesis.
Comparison with other methods and additional applications
As discussed above, the major difference between the iG model and
other models is that the iG model achieves an absolute control over
the intragastric delivery of nutrients or any liquid test substance.
It allows a precise control over the feeding or administration rate
(amount) and mode (continuous, intermittent or bolus). More
importantly, it allows testing the role of a segment of the normal
oroesophageal and GI route of nutrient intake by bypassing it via
intragastric or intraintestinal access. Two limitations of the model
are (i) that the diet or substance to be administered must be in a
liquid form, and (ii) that it requires surgical expertise and swivel
and infusion setup for chronic infusion.
Although this article describes technical details of the iG mod-
eling in the mouse to help readers acquire essential methodologi-
cal information, the same principle can be applied for the rat as
described elsewhere26. Dietary and infusion methods described
pertain to continuous intragastric infusion of a liquid diet with a
varying content of corn oil with or without ethanol administration
as an example of the iG model application, but they can be modified
according to different experimental needs.
Animals. Any strain, age or sex of mice can be used for the model.
Genetically modified mice or mice subjected to nutritional or phar-
macological treatments may differ in their response to dietary or
ethanol administration. For the standard mouse iG model, we
use 8-week-old male C57BL/6J mice obtained from the Jackson
Laboratory. We allow at least 1 week of acclimatization after receiv-
ing mice before the iG surgery. Another week of a recovery period
follows the surgery, and during this period a basic control diet is
infused with or without regular chow.
Controls. Appropriate controls should be created for an experi-
mental group, using identical strains, ages, sexes and iG feeding
procedures. It is ideal to use the same littermates. For ethanol
experiments, an isocaloric diet replacing calories from ethanol with
dextrose solution can be used to pair feed. To examine the effects of
iG overfeeding, mice fed at the standard caloric intake can be used
as a control. In addition, if the amount of a nutrient is modified,
the unmodified diet can be used for controls.
Mice: any strain, age or sex can be used ! cautIon Experiments involving
mice must conform to governmental and institutional ethics regulations.
Gastrostomy catheter reagents
Chloroform (Sigma, cat. no. C2432) crItIcal Always use chloroform
fresh from the bottle.
Anesthesia and pre-/postoperative reagents
Ketamine HCl Injection, USP (Western Medical, cat. no. NDC-57319-542-02)
Xylazine sterile solution (Western Medical, cat. no. NADA no. 139-236)
Gentamicin sulfate veterinary ophthalmic ointment (Western Medical,
cat. no. NDC17033-011-38)
Buprenorphine (Buprenex) hydrochloride injection (Western Medical,
cat. no. NDC 40042-010-01)
Prodine (povidone-iodine) solution (Western Medical,
cat. no. NDC57319-328-09)
Lactated ringer’s and 5% dextrose injection (Western Medical,
cat. no. NDC0409-7929-09)
Ethanol (70%, vol/vol; diluted from 100% solution)
Sodium chloride irrigation (0.9%), USP (Western Medical,
cat. no. NDC0409-7138-09)
Lactalbumin hydrolysate (Invitrogen, cat. no. 11800-042)
Dextrose (Sigma, cat. no. G8270)
Citric acid (Sigma, cat. no. C0759)
Potassium phosphate (Sigma, cat. no. P5379)
Calcium chloride (Sigma, cat. no. C8106)
Sodium chloride (cat. no. S3014)
Magnesium sulfate (Sigma, cat. no. M1880)
Manganese sulfate (Sigma, cat. no. M7634)
Potassium iodide (Sigma, cat. no. P207969)
Ammonium molybdate (Sigma, cat. no. A7302)
Cupric sulfate (Sigma, cat. no. C7631)
Ferric ammonium citrate (cat. no. F5879)
Zinc chloride (Sigma, cat. no. Z208086)•
Trace mineral mix (Dyets, cat. no. 210090) crItIcal A fresh batch
should be aliquotted into small vials and stored in a refrigerator. Take out
one vial for each use as needed.
Vitamin mix AIN-76 (Dyets, cat. no. 300050) crItIcal Aliquots should
be prepared in a manner similar to that mentioned above.
Choline chloride (Dyets, cat. no. 400775) crItIcal Aliquots should be
prepared in a manner similar to that mentioned above.
Mazola corn oil or any other fat type (Local grocery store)
Xanthan gum (Sigma, cat. no. G1253)
Ethanol (100%, vol/vol; Fisher, cat. no. 04-355-451)
Gastrostomy catheter materials
Tygon tubing (inner diameter (i.d.) 0.508 mm × outer diamter (o.d.) 1.524 mm;
i.d. 0.762 mm × o.d. 2.286 mm; Fisher, cat. nos. 1417015B, 1417015C)
Silastic silicone tubing (i.d. 0.635 mm × o.d. 1.1938 mm, i.d. 0.762 mm ×
o.d.1.651 mm; Fisher, cat. nos. 11-189-15B, 11-189-15C)
Stainless steel wire T-340 V (Small Parts, cat. no. GWX-0110-30)
Dacron felt (0.635 mm in thickness, PEI)
RTV silicone rubber (Ellsworth, cat. no. 2118874)
RTV catalyst no. 4 (Ellsworth, cat. no. 1103181)
Precision swivel (22 G; Instech, cat. no. 375-22P)
Intramedic Luer stub adaptor (23 G; VWR, cat. no. 63019-852)
Intramedic Luer stub adaptor (20 G; VWR, cat. no. 63019-841)
Syringe (1 ml; VWR, cat. no. BD309602)
Y-shaped tubing connector (Small Parts, cat. no. STCY-18-10)
Nylon tubing (Small Parts, cat no. NSR-106B-C)
6/0 silk suture (Hospital Associates, cat. no. 1639G)
Powdered sterile latex surgical gloves (Fisher, cat. no. 19-020-557B)
Eye dressing surgical forceps (101.6 mm; VWR, cat. no. 21909-694)
Nonsterile cotton gauze sponges (100 mm × 100 mm; Fisher,
cat. no. MSD14-002-50)
Surgical towels (457.2 mm × 838.2 mm)
Sterilization pouch (190.5 mm × 330.2 mm; VWR, cat. no. 89140-802)
Castroviejo needle holder (139.7 mm; VWR, cat. no. 10049-416)•
© 2012 Nature America, Inc. All rights reserved.
nature protocols | VOL.7 NO.4 | 2012 | 773
Jacobson Mosquito surgical forceps straight, curved (VWR, cat. nos.
Miltex Jeweler-style forceps (Express Surgical (http://www.expresssurgical.com/),
cat. no. 17-305)
Water heating pump (Fisher, cat. no. NC9803964)
Spectra mesh filter (70 µm; Spectrum lab, cat. no. 146-490)
Sterile serological pipette (25 ml)
Glass beakers (3 liters)
Plastic beakers (4 liters)
Graduated cylinders (2 liters)
Graduated cylinders (1 liter)
Stirrer and stirring magnetic bars
VLP 200 mechanical vacuum pump (Ideal vacuum products, cat. no. VLP200)
Plastic Buchner filtering funnel•
Commercial food blender (3 liters; Pacific Kitchen Equipment,
cat. no. 38BL61)
Glass filtering flask (2 liters)
Infusion setup supplies
Syringe (10 ml; VWR, cat. no. BD309604)
Shoe-box mouse cage with filter top lid
Stainless steel rack for animal cages with steel rod for clamps
Rubber stopper with a hole, cut to hold swivel
Harvard Apparatus PHD 2000 Infusion pump with multirack kit
(Harvard Apparatus, cat. no. PHD70-2003)
Water-circulating heating pad
Silastic catheter tip•
1| Cut a Tygon tube (i.d. 0. 508 mm × o.d. 1.524 mm; ‘b’ in Fig. 1) to a length of 350 mm.
2| Cut a Silastic tube (i.d. 0.64 mm × o.d. 1.19 mm; ‘f’ in Fig. 1a) to a length of 70 mm and immerse it in chloroform for a
few minutes to expand the diameter.
! cautIon This should be done in a fume hood.
3| Slide the Silastic tube over the end of the Tygon tube up to about 25 mm.
crItIcal step Hold the Silastic tube in place for 30 s to prevent it from slipping off of the Tygon tube.
4| Measure and mark (with a marker pen) a distance of 10 mm from the end of Tygon tube with the Silastic tube over it.
Bend the tubing at this precise location and hold this bend by placing it into rubber tubing (i.d. 12 mm × o.d. 17 mm in a
10-mm length). Place the angled area supported by the rubber tubing into boiling water for about 3 s and remove the rubber
tube from the angle point. This will create a 90–100° angle (Fig. 1c).
! cautIon Be careful when handling boiling water and the hot plate.
crItIcal step Do not make too sharp of an angle, as it narrows the diameter of the Tygon tubing.
5| Attach a stainless steel wire
(150 mm in length; ‘d’ in Fig. 1a) along
the outer wall of the proximal Tygon tub-
ing by securing it with short, chloroform-
expanded Silastic tubing at four or five
positions on the catheter (‘c’ in Fig. 1a).
6| Cut a piece of Dacron felt to
a pear shape (8 mm width, 16 mm
length) and pass the catheter through
a hole made in the middle (‘e’ in
Fig. 1a); next, glue it to the catheter
angle with RTV silicone rubber resin
and catalyst (Fig. 1c,d).
7| Cut a Dacron felt disc (6 mm in
diameter), pass the Silastic catheter
tip via a hole made in the middle and
glue the disc to Silastic tubing at a
location 33 mm distal to the tip of the
Tygon tubing (‘g’ in Fig. 1a).
Figure 1 | Catheter setup. (a) A mouse catheter diagram showing different components: a, swivel;
b, Tygon tube; c, Silastic tube; d, stiffening wire; e, pear-shaped piece of Dacron; f, Silastic tube and
g, Dacron disc. (b) An overall photo of the catheter. (c) Close-up side view of the catheter tip. (d) Close-up
bottom view of the catheter tip.
© 2012 Nature America, Inc. All rights reserved.
774 | VOL.7 NO.4 | 2012 | nature protocols
8| Place a 1-ml syringe with a 23-G Luer stub adaptor on the proximal end of the catheter (Fig. 1b).
9| Gas-sterilize the catheter in a sterilization pouch before its use.
pause poInt The catheter can be presterilized and stored in a sealed pouch for 3–4 weeks.
surgical procedure ● tIMInG 40 min maximum
10| Anesthetize a mouse (body weight 24–27 g) by intraperitoneal injection of ketamine (80 mg kg − 1) and xylazine
(10 mg kg − 1). Ketamine and xylazine should be premixed and diluted with saline.
crItIcal step The surgery procedure should be complete in ~40 min or less. Anesthesia achieved by this regimen should
sufficiently cover this duration.
11| Preparation of the mouse for aseptic surgery. Clip the hair on the mid-abdomen and dorsal neck areas. Swab them with
iodine solution and allow them to dry. Pull out the tongue and apply ointment on the mouse′s eyes. Wipe the surgical areas
with 70% (vol/vol) ethanol. Place the mouse on a surgical table that has been covered with a sterile drape and heated with
a water-circulating heating pad (set at 37 °C) placed underneath the drape to prevent hypothermia. Use sterile gauze to
cover areas that have not been prepared. Fill a catheter with saline using a 1-ml syringe.
crItIcal step Sterilized surgical instruments, towels, gowns, catheter, gloves and sutures must be used.
12| Place the mouse on its stomach. Make a skin incision (8–10 mm in length) on the dorsal midline from the base of the
skull to the interscapular region (Fig. 2a). Use scissors to separate the skin from the underlying muscle tissue around the
incision site and extending toward the left flank; this will create a subcutaneous tunnel.
13| Position the mouse on its back. Make a skin incision (8–10 mm) on the abdominal midline from the xiphoid cartilage
extending to the midabdomen (Fig. 2b). Use scissors to separate the skin from the underlying muscle tissue. This separation
process will move toward the left flank to complete a connection with the subcutaneous tunnel created from the neck.
14| Position the mouse on its right side. Insert a curved hemostat from the abdominal skin incision into the subcutaneous
tunnel and out of the dorsal neck incision (Fig. 2c). By using the hemostat, gently grasp the catheter above the Dacron disc
(Fig. 2d). Carefully pull the catheter tip through the subcutaneous tunnel out of the abdominal incision (Fig. 2e).
crItIcal step Do not grasp the Dacron disc. The integrity of the disc and its adhesion to the tubing may be
Figure 2 | Surgical procedure for implantation
of a gastric catheter. (a) Dorsal midline incision
of the neck. (b) Ventral abdominal midline
incision. The skin layer has been separated
from the underlying abdominal muscle.
(c) A curved hemostat is inserted through the
ventral incision out of the neck incision via a
subcutaneous tunnel. (d) The hemostat grasping
the catheter above the Dacron (stomach) disc.
(e) The catheter is pulled out of the ventral
incision. (f) The peritoneum is opened.
(g) A curved hemostat is inserted into the
abdomen, placed against the inner wall of the
abdomen and forced to penetrate the wall at
a distance of 20 mm away from the midline in
the left flank area. The hemostat grasping the
catheter behind the Dacron (stomach) disc.
(h) The catheter is pulled into the cavity and
out of the midline incision. (i) The catheter
tip is positioned near the area of insertion site
of exposed forestomach. (j) A puncture hole is
made in the middle of the forestomach, and the
catheter tip is inserted into the forestomach.
(k) The Dacron (stomach) disc is sutured to the
forestomach wall by placing 4–6 stitches
with 6/0 silk. (l) Closing of the abdominal
incision with 6/0 silk. (m) The ventral skin incision is closed with 6/0 silk. (n) Dorsal Dacron is positioned along the cervical muscles within the skin
incision. (o) Dacron is sutured to cervical muscles. (p) The skin incision is closed for the catheter exit site.
© 2012 Nature America, Inc. All rights reserved.
nature protocols | VOL.7 NO.4 | 2012 | 775
15| Make a midline incision on the linea alba to open the abdominal cavity (6–8 mm; Fig. 2f).
16| Insert a pointed curved hemostat into the abdomen, place the tip against the inner abdominal wall at the site ~20 mm
toward the left flank from the incision, and then make a hole in the abdominal wall by pressuring the tip and penetrating the
wall (use a scalpel tip to assist if needed). By using the hemostat, grasp and pull the catheter tip into the abdomen (Fig. 2g,h).
17| Use a curved forceps to find the forestomach by flipping the left liver lobe. Be careful not to damage the liver or any
other major organs. The forestomach can be distinguished from the glandular portion of the stomach, as it is whiter in color.
Gently grasp the forestomach and pull it out of the incision site (Fig. 2i).
! cautIon Be careful not to pinch and damage blood vessels on the forestomach. Place the stomach on gauze soaked
18| Use a jeweler′s forceps to make a small puncture hole (a size of the diameter of Silastic tube) through the forestomach
wall in the middle of the forestomach where there is minimal vasculature. Insert the catheter tip into forestomach. With
curved forceps, hold the catheter in place (Fig. 2j). Use saline to prevent dehydration of the exposed stomach. Catheter tip
placement and subsequent suturing are relatively easier in the morning when the mouse stomach is still full of diet
consumed the previous night. As gastric emptying becomes more complete in the afternoon, gastric liquid content
tends to spill out of the puncture site. Try to block and absorb this with gauze.
19| Use 6/0 silk or other nonabsorbable suture material to stitch the Dacron disc to the forestomach wall. Four to six
stitches around the disc perimeter will secure its placement (Fig. 2k). After completing the suturing, gently infuse ~0.1 ml
saline into the stomach using the saline-filled syringe attached to the proximal end of the catheter. Ensure that there is no
leakage from the sutured disc.
20| Replace the whole stomach to the original resting position by pushing it into the abdomen while gently pulling the
catheter from the dorsal neck side. The Dacron disc should be placed against the abdominal hole in the left flank.
! cautIon Ensure that the spleen and a splenic portion of pancreas are not trapped by the disc.
21| Close the peritoneal cavity using 6/0 silk or absorbable suture and a continuous suture pattern (Fig. 2l).
! cautIon Be careful not to puncture the internal organs, including abdominal fat.
22| Close the abdominal skin incision using 6/0 silk or other nonabsorbable suture and an interrupted suture pattern (Fig. 2m).
23| Position the mouse on its stomach. Position the anchoring Dacron along the dorsal neck muscle (Fig. 2n).
24| Use 6/0 silk to suture the Dacron onto the neck muscle by placing two stitches on either side (Fig. 2o).
! cautIon This suture is crucial to firmly anchor the catheter entry site. Ensure that you stitch a sufficiently deep layer of
the muscle to achieve this.
Figure 3 | Mice during the experiment. (a) An iG mouse inside the cage. (b) Diagram depicting how infusion lines are connected to a swivel: a, swivel;
b, Silastic tube; c, Y-shaped tubing connector; d, Tygon tube; and e, 20-G Intramedic Luer stub adaptor. (c) Overview photo showing a setup for cages and
infusion pumps on racks, as well as swivels supported by clamps and rods attached to the rack.